vault backup: 2024-09-29 16:20:52

Record/DL/CoT Prompt.md

@@ -1,17 +1,11 @@
-## Standard Prompt
-
-> **Model Input:** Q: Roger has 5 tennis balls. He buys 2 more cans of tennis balls. Each can has 3 tennis balls. How many tennis balls does he have now?
-> A: The answer is 11.
-> Q: The cafeteria had 23 apples. If they used 20 to make lunch and bought 6 more, how many apples do they have?
-> **Model Output:** A: The answer is 27.（x）
-
-## Few-shot-CoT
-Prompt: `rationale(r) + answer(a)`
-Answer: LLM同样会给出理由和答案。
-
-> **Model Input:** Q: Roger has 5 tennis balls. He buys 2 more cans of tennis balls. Each can has 3 tennis balls. How many tennis balls does he have now?
-> A: Roger started with 5 balls. 2 cans of 3 tennis balls each is 6 tennis balls. 5 + 6 = 11. The answer is 11.
-> Q: The cafeteria had 23 apples. If they used 20 to make lunch and bought 6 more, how many apples do they have?
-> **Model Output:** A: The cafeteria had 23 apples originally. They used 20 to make lunch. So they had 23 - 20 = 3. They bought 6 more apples, so they have 3 + 6 = 9. The answer is 9.
-
+## Standard Few-shot Prompt
+Prompt: `Q(question) + A(answer)`
+## Few-shot CoT
 思维链提示，就是把一个多步骤推理问题，分解成很多个中间步骤，分配给更多的计算量，生成更多的 token，再把这些答案拼接在一起进行求解。
+Prompt: Q + A(r(rationale) + a(answer))
+Answer: LLM同样会给出理由和答案。
+## Zero-shot CoT
+通过 `Let's think step by step` 可以让 LLM 生成回答问题的思维链。我们可以将 Zero-shot CoT 看作一个 pipeline，我们先使用  `Let's think step by step` 让 LLM 尽可能生成一些思考过程，然后将生成的 rationale 和 question 拼接起来，重新配合一个指向 answer 的 Prompt 来激励模型生成答案。
+Output: (Q + (Q + Let's think step by step ｜ LLM) + The answer is) | LLM
+## Self-consistency
+自洽性，使用多数投票(majority vote)的方式来提升最终回答的准确性。不同于之前的方法，本方法会生成多个思路链，最后取多数答案作为最终答案。